Chemical and Pharmaceutical Bulletin
Online ISSN : 1347-5223
Print ISSN : 0009-2363
ISSN-L : 0009-2363
Regular Articles
Molecular Orbital Study of the Formation of Intramolecular Hydrogen Bonding of a Ligand Molecule in a Protein Aromatic Hydrophobic Pocket
Jun KosekiHiroaki GoudaShuichi Hirono
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML
Supplementary material

2016 Volume 64 Issue 7 Pages 1031-1035

Details
Abstract

The natural product argadin is a cyclopentapeptide chitinase inhibitor that binds to chitinase B (ChiB) from the pathogenic bacteria Serratia marcescens. Nω-Acetyl-L-arginine and L-aminoadipic acid of argadin form intramolecular ionic hydrogen bonds in the aromatic hydrophobic pocket of ChiB. We performed ab initio molecular orbital and density functional theory calculations to elucidate the role of this intramolecular hydrogen bonding on intermolecular interactions between argadin and ChiB. We found that argadin accrues large stabilization energies from the van der Waals dispersion interactions, such as CH–π, π–π, and π–lone pair interactions, in the aromatic hydrophobic pocket of ChiB, although intramolecular hydrogen bonding within argadin might result in loss of entropy. The intramolecular ionic hydrogen bonding formation canceled local molecular charges and provided good van der Waals interactions with surrounding aromatic residues.

Ligand–protein interactions result from a combination of effects, for example, hydrogen bonding and van der Waals interactions such as CH–π and π–π dispersion interactions. In general, a ligand-binding site has hydrophobic areas consisting of non-polar residues and hydrophilic areas consisting of polar residues. The hydrogen bonding regions of the ligand are located near polar residues or water molecules, and the hydrophobic regions are covered with non-polar residues, providing the stabilization energy for binding to the ligand-binding site.

Recently, Houston et al.1,2) reported X-ray crystal structure analysis of the binding conformation of the natural product argadin (Fig. 1), which is a cyclopentapeptide chitinase inhibitor, in the binding site of Serratia marcescens chitinase B (ChiB). The binding site of ChiB has an aromatic hydrophobic moiety consisting of tryptophan (Trp)97, Tyr145, phenylalanine (Phe) 191, and Trp220. At this hydrophobic moiety, two charged amino acids of argadin, L-aminoadipic acid (negative) and Nω-acetyl-L-arginine (positive), form a double hydrogen bonding system (Fig. 2A). These two amino acids are sandwiched between Trp97 and Trp220 (Fig. 2B). The intramolecular double hydrogen bonding region is located close to the aromatic residue Trp220. It is rare for the charges in ligand molecules to be canceled by forming intramolecular double hydrogen bonds. In a previous computational study, we performed molecular dynamics simulations of the argadin–ChiB complex and free argadin to show that the intramolecular double hydrogen bonding of argadin was stable in the ChiB-bound state, but it was rarely observed in the free state.3) Therefore, the formation of intramolecular hydrogen bonding inflicts a large loss of entropy when argadin binds to ChiB, although this loss is compensated for by sufficient stabilization energy upon argadin–ChiB complex formation. However, the size of the stabilization energy that argadin gains in the hydrophobic binding site of ChiB and the type of interaction between the surrounding residues and the intramolecular double hydrogen bonding system are unknown.

Fig. 1. Two-Dimensional Structure of Argadin

L-Aminoadipic acid and Nω-acetyl-L-arginine, which form an intramolecular ionic hydrogen bond in the hydrophobic pocket of ChiB, are labeled.

Fig. 2. (A) Conformation of Argadin Bound to ChiB; (B) Positional Relationship of Trp97 and Trp220 for Argadin

(A) The aromatic hydrophobic pocket of ChiB consisted of Trp97, Tyr145, Phe191, and Trp220. (B) Argadin is sandwiched between these two tryptophans.

In this study, we performed ab initio molecular orbital calculations including the contribution of the dispersion interaction to qualitatively reveal the role of the intramolecular double hydrogen bonding system in the hydrophobic pocket. The details of the analysis of the intermolecular interactions with some fragment-models are shown in Experimental. The calculated interaction energies between argadin and each aromatic hydrophobic residue and the role of double hydrogen bonding system are discussed in Results and Discussion. Finally, concluding remarks are given in Conclusion.

Experimental

To calculate the stabilization energies and analyze the donor-acceptor interactions between argadin and each aromatic hydrophobic residue, we performed ab initio molecular orbital and density functional theory calculations. First, the whole ChiB–argadin complex structure was relaxed in the water solvent by using an energy minimization method with the AMBER99 force field46) in AMBER9.7) In this minimization calculation, we used the ChiB–argadin complex taken from the crystal structure (PDB ID: 1H0G1)) as the initial coordinates. Two levels were used to optimize structures with the ONIOM method.8) For the first level, we defined argadin as the high-level layer with B3LYP/6-31G*, and the all atoms of ChiB as the low-level layer with the AMBER99 force field.4) For the second-level calculation, we defined the argadin and four aromatic hydrophobic residues (Trp97, Tyr145, Phe191, and Trp220) as the high-level layer with Møllor–Plesset perturbation method (MP2)/STO-3G, and the other residues of ChiB with the AMBER99 force field in parallel.4) In MP2 calculation, unfortunately, using the minimum basis set was limited by our computational resource although we should ideally use more high level basis set. During these optimized calculations, all atoms assigned as the low-level layer were restricted to the position of the first minimization coordinate. From these optimized structures, the model structures were constructed from argadin and four aromatic hydrophobic residues. The α-carbon of each residue was terminated by a methyl group in our model system. The model structures optimized at the B3LYP/6–31G level and at the MP2/STO-3G level were named ST1 and ST2, respectively. Calculating the protein–ligand interaction energy for the four hydrophobic aromatic residues with argadin in our model complex system, we estimated the approximate stabilization energy ∆SE as   

(1)
where Ecomplex, Earg, and E4arom are the total energy of the ChiB–argadin complex (four hydrophobic aromatic residues and argadin), argadin, and the four aromatic hydrophobic residues, respectively. In addition, to analyze the detailed interaction energies between argadin and each aromatic residue, we estimated the two-body interaction energies, ∆IEi, as   
(2)
where Earg+i_arom is the total energy of argadin and the hydrophobic aromatic residues, and Ei_arom is the total energy of the aromatic residues. In these calculations, the stabilization energies and interaction energies were calculated with the Hartree–Fock (HF) method and second-order Møllor–Plesset perturbation method (MP2) to reveal the characteristics of the intermolecular interactions between argadin and the aromatic residues. The HF energy includes the contribution of dispersion interaction, whereas the MP2 energy does not. We performed natural bond orbital (NBO) analysis9) for the complex of argadin and four hydrophobic aromatic residues to reveal the donor–acceptor interaction between argadin and these residues. The electronic donor–acceptor interaction energies for Lewis type NBO, ∆Eij, were calculated with Eq. (3).   
(3)
These interaction energies are the stabilization energies associated with electron transfer between specified NBO i (Donor) and j (Acceptor) having corresponding to orbital energies εi and εj, and are described as the second order perturbation energies for Fock operator F.10) The 6-31G* basis set was used to calculate stabilization energies and ∆Eij for the NBO analysis. All ab initio molecular orbital calculation were performed with Gaussian 03.11)

Results and Discussion

Table 1 shows the stabilization energies, ∆SE, between argadin and hydrophobic aromatic residues for the ST1 and ST2 model structures calculated with HF and MP2 levels using the 6–31G* basis set. The stabilization energies for ST1 and ST2 with MP2 calculations are −28.59 and −31.18 kcal/mol, respectively, whereas the HF calculations give positive values of 9.72 and 3.59 kcal/mol, respectively. These results clearly show that the van der Waals dispersion interactions are an important factor in the complex formation because the contribution of the dispersion force is not included in the HF calculation, whereas the effect of the dispersion interaction can be included in the calculation with the MP2 approximation. Table 1 also shows the interaction energies, ∆IE, between argadin and Trp97, Tyr145, Phe191, and Trp220. The interaction energies, ∆IEW97, ∆IEY145, ∆IEF191, and ∆IEW220, are −14.11 (−16.86), −1.18 (−0.65), −3.85 (−3.39), and −9.32 (−10.19)  kcal/mol, respectively, for ST1 (ST2) calculated with MP2. The sum of these interaction energies for each model structure with MP2 calculations are −28.46 and −31.09 kcal/mol, respectively, which are in reasonable agreement with the corresponding stabilization energies of −28.59 and −31.18 kcal/mol. Therefore, the interactions involved in stabilizing argadin in the aromatic hydrophobic binding pocket of ChiB can be separated into the two-body interaction energies, ∆IEs, and the three-body interactions can be ignored. ∆IEW97 and ∆IEW220 are the main interaction energies, because the sum of these two energies accounts for 82% (87%) of the stabilization energy for ST1 (ST2). The interaction energies estimated with ST1 and ST2 are similar, because the structures in ST1 and ST2 are similar (Fig. 3). Though the moieties of Pro in Argadin between ST1 and ST2 might have a little difference of their orientation, the behavior arose from existing in solvent accessible area. Then, it is not essential in this study. This result means that the conformation of Trp97 and Trp220 for argadin represented with the energy minimization method with the AMBER99 force field was as good as the structure optimized with the MP2 level.

Table 1. Stabilization Energies (kcal/mol) between Argadin and Hydrophobic Aromatic Residues with HF, MP2 Levels Calculated by Using the 6–31G* Basis Set
–ST1–SEIEW97IEY145IEF191IEW220
HF9.725.99−0.36−0.504.56
MP2−28.59−14.11−1.18−3.85−9.32
–ST2–SEIEW97IEY145IEF191IEW220
HF3.590.990.29−0.873.15
MP2−31.18−16.86−0.65−3.39−10.19

The energies decomposed into the interaction energies of the factors of each hydrophobic aromatic residue of Trp97, Tyr145, Phe191, and Trp220, are also shown.

Fig. 3. The Comparison of Optimized Structures, ST1 and ST2

ST1 and ST2 are the optimized structures at the B3LYP/6–31G level for argadin, and with MP2/STO-3G level for argadin, Trp97, Tyr145, Phe191, and Trp220, respectively, after energy minimization with the AMBER99 force field.

Argadin is sandwiched between Trp97 and Trp220 in the aromatic hydrophobic pocket of ChiB. Trp220 is located close to the intramolecular double hydrogen bonding of argadin (Fig. 2B). We examined the large interaction energies of argadin with Trp97 and Trp220 in the aromatic hydrophobic pocket. To reveal the detail of the interactions between argadin and these residues, NBO analysis was performed. We used representative CH–lone pair (LP), CH–π, π–π, and π–LP interactions from the NBO analysis between Trp97 and argadin NBO and between Trp220 and argadin NBO (Fig. 4). The electronic donor–acceptor interaction energies are shown in Tables 2 and 3, respectively. In these tables, BD and LP mean bonding orbital and lone pair, where the orbitals adding “*” mean non-Lewis orbitals. Some of these interactions could be explained as a van der Waals interaction. First, we focus on the interaction between Trp97 and argadin. There are many donor-acceptor energies caused by CH–π interactions (Fig. 5). In Fig. 5, the representative CH–π interaction between the π-bonding orbital of C17–C19 and the antibonding orbital of C161–H162 is shown. The interaction energies caused by these CH–π interactions between C17–C19 and C161–H162, between C7–C8 and C84–H85, and between C152–153 and C4–H5 are greater, and their values are −1.81, −1.59, and −1.34 kcal/mol, respectively. These results show that the stabilization factors between Trp97 and argadin are mainly caused by these CH–π interactions, although there may be other factors such as CH–LP interactions. Next, we focus on the interactions between Trp220 and argadin. Some π–π and π–LP interactions are observed in addition to the CH–π interactions. Figure 6 shows a representative π–LP interaction between the π–bonding orbital of C72–C74 and the lone pair of C167 (upper), and the π–π interaction between the π–bonding orbital of C76–C78 and the π*–antibonding orbital of N95–C97 (lower). The interaction energies of these π–π and π–LP interactions are −1.13 and −0.11 kcal/mol, respectively. There are many interactions with the π–conjugate bonding region of the double hydrogen bonding system similar to the interactions between aromatic rings, although each interaction energy between Trp220 and argadin is weak compared with that between Trp97 and argadin. Therefore, the molecular orbitals of argadin were checked with the MP2/6–31G* level. We observed a conjugate-like orbital in the double hydrogen bonding system between Nω-acetyl-L-arginine and L-aminoadipic acid in argadin (Fig. 7). These orbitals are similar to the π and π* orbitals of benzene. These results can be explained by the intramolecular double hydrogen bonding cancelling local molecular charges and the formation of good van der Waals interactions with the surrounding hydrophobic aromatic residues, resulting in the large stabilization energy of the ChiB–argadin complex.

Fig. 4. Numbering Scheme for Trp97, Trp220, and Argadin
Fig. 5. Representative NBO of the CH–π Interaction between the π–Bonding Orbital of C17–C19 and the Antibonding Orbital of C161–H162
Fig. 6. Representative NBO of the π–LP Interaction between the π–Bonding Orbital of C72–C74 and the Lone Pair of C167 (Upper), and π–π Interaction between the π–Bonding Orbital of C76–C78 and the π*-Antibonding Orbital of N95–C97 (Lower)
Fig. 7. Conjugate-Like Orbital in the Double Hydrogen Bonding System between L-Aminoadipic Acid and Nω-Acetyl-L-arginine of Argadin (Left Side), and the Corresponding π–Conjugate Orbital of Benzene (Right Side)
Table 2. Interaction Energies (kcal/mol) between Trp97 and Argadin NBO
(A) From Trp97 to argadin NBO
Donor NBOAcceptor NBOEnergy
BDC4–H5BD*C152–O153−0.11
BDC7–C8BD*C84–H85−1.59
BDC7–C8BD*C120–C122−0.62
BDC12–C21BD*C89–H90−0.06
BDC12–C21BD*C89–H91−0.06
BDC17C19BD*C156–H157−0.10
BDC17–C19BD*C161–H162−1.81
LPN10BD*C89–H90−0.42
BD*C7–C8BD*C120–H122−0.33
BD*C7–C8BD*O170–C171−0.39
BD*C12–C21BD*C89–H90−0.18
BD*C12–C21BD*O170–C171−0.07
BD*C17–C19BD*C156–H157−0.14
(B) From argadin to Trp97 NBO
Donor NBOAcceptor NBOEnergy
BDC89–H90BD*N10–C12−0.05
BDC89–H90BD*C12–C21−0.33
BDC89–H90BD*C13–C15−0.11
BDC120–H122BD*C7–C8−0.17
BDC152–O153BD*C4–H5−1.34
BDC156–H157BD*C17–C19−0.05
BDO170–C171BD*C4–H5−0.19
LPO153BD*C4–H5−1.46
LPO153BD*C4–H5−0.58
LP*C167BD*C17–C19−0.09
BD*C152–O153BD*C4–H5−0.14
The bonding types are provided in Supplementary Table 1.
Table 3. Interaction Energies (kcal/mol) between Trp220 and Argadin NBO
(A) From Trp220 to argadin NBO
Donor NBOAcceptor NBOEnergy
BDC63–H65BD*C103–O104−0.16
BDC72–C74LP*C167−1.13
BDC72–C74BD*C164–H166−0.33
BDC76–C78LP*C167−0.22
BDC76–C78BD*N95–C97−0.11
BDC78–H79BD*N95–C97−0.07
(B) From argadin to Trp220 NBO
Donor NBOAcceptor NBOEnergy
LP*C167BD*C72–C74−0.30
LP*C167BD*C76–C78−0.10
LPO168BD*C72–C74−0.08
LPO168BD*C76–C78−0.59
BD*N95–C97BD*C76–C78−0.24
BD*C103–O104BD*C66–C67−0.05
The bonding types are provided in Supplementary Table 1.

Conclusion

We have performed molecular orbital study to elucidate the role of intramolecular hydrogen bonding within argadin in the intermolecular interactions at the hydrophobic binding pocket in ChiB, consisting of four hydrophobic aromatic residues, Trp97, Tyr145, Phe191, and Trp220. Argadin accrues large stabilization energies from the van der Waals dispersion interactions, such as CH–π, π–π and π–LP interactions, at the ChiB hydrophobic pocket, although the intramolecular hydrogen bonding of argadin may result in a loss of entropy. The formation of intramolecular double hydrogen bonding may cancel local molecular charges and provide good van der Waals interactions with the surrounding hydrophobic aromatic residues.

Acknowledgment

Our calculations were performed mainly using the computational resources of the Quantum Chemistry Division at Yokohama City University.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

The online version of this article contains supplementary materials.

References
 
© 2016 The Pharmaceutical Society of Japan
feedback
Top